NKI-Rockland Sample (pilot, Phase I)

The NKI-Rockland Sample (NKI-RS) is a discovery-science project of the Nathan Kline Institute for Psychiatric Research in Rockland County, New York. Its pilot phase collected brain MRI together with a broad set of psychiatric, cognitive and behavioral measures from people across the lifespan, and released the data through the International Neuroimaging Data-sharing Initiative while recruitment was still running. Weekly sharing started in October 2010 and the last release set was completed on August 15, 2011. This entry covers that pilot release. The later Enhanced NKI-RS (Phase II, from March 2012) is a separate, community-ascertained sample that requires a data usage agreement and is not part of this entry.

Composition

The release table on the project page lists eight sets with 20 to 30 people each, 207 in total. The data descriptor gives a higher figure: 250 individuals collected and released in 11 months. The project page describes the sample as aged 4 to 85; the pilot targeted ages 4 to 89. Recruitment was by convenience, open to anyone willing to take part in that age range. Participants had semi-structured diagnostic interviews, and more than 300 phenotypic variables from 26 domains were gathered. A combined clinical information file accompanies the images.

Acquisition

All published protocols ran on a Siemens MAGNETOM TrioTim and were adapted from the Brain Genomics Superstruct Project. They comprise a 10-minute resting-state fMRI run (3 mm isotropic, TR 2.5 s), a 6-direction and a 64-direction diffusion scan (the latter 2 mm isotropic), a standard and a shorter accelerated MPRAGE (1 mm isotropic), and a short sequence the page calls T2-weighted, which the protocol sheet shows as a single-volume EPI. Images are offered as DICOM or NIfTI, defaced and with randomized release timing.

Annotations

There are no image labels. Phenotypic data are summary scores from the assessment battery.

Known limitations

• Convenience sampling may not represent the local population.
• Phenotypic measures were chosen for availability rather than as normed instruments, and only summary scores were kept.
• DKEFS scores of 36 subjects were rescored after release; earlier downloads need the DKEFSfix update.
• Scans failing quality control are flagged but still released.
• Defacing can disturb some analysis tools, such as FreeSurfer.
• Subject counts differ between the release table and the paper.